Statistical evaluation was performed either by Student's t-test (two-tailed) or by Mann–Whitney U-test (two-tailed) for means between patient and control groups, or by Pearson's method for correlation, using SPSS software version 19 (IBM, Tokyo, Japan). Within each group, the subjects with transcript levels >2 s.d. from the group mean were considered as outliers and removed. In addition, for the expression analyses of GFAP, MAP2 and ratio of GFAP/MAP2, the P-values were also calculated using analysis of covariance by adjusting the age.
